Month: May 2016

ARTO

Time for a sta­tus update on my stay in Sin­ga­pore. I have already entered the final three months of my time here. Time flies when you’re hav­ing fun eat­ing every­thing in sight, it turns out. 

On the work front I have indeed found the time to do some think­ing about what my next big thing will be. Noth­ing has firmed up to the point where I feel like shar­ing it here but I am enjoy­ing the con­ver­sa­tions I am hav­ing with var­i­ous peo­ple about it.

In the mean­time, I have been keep­ing busy work­ing with a local start­up called ARTO. I have tak­en on the role of prod­uct design­er and I am also respon­si­ble for prod­uct man­age­ment of the user-fac­ing parts of the thing we are building. 

That “thing” is about art. There are many peo­ple who are inter­est­ed in art but don’t know where to start when it comes to find­ing, enjoy­ing and acquir­ing it. We’re build­ing a mobile and TV app that should make that a whole lot more easy and fun.

When I say art I mean com­mer­cial, pop­u­lar and con­tem­po­rary art of the 2D vari­ety. So paint­ing, illus­tra­tion, pho­tog­ra­phy, etc. Things you might buy orig­i­nals or prints of and put on your liv­ing room wall. Oth­ers are doing a fine job on the high end of the art mar­ket. We think there are parts remain­ing that have been under­served to date. 

There are many mov­ing parts to this prod­uct, rang­ing from a rec­om­men­da­tion engine, con­tent man­age­ment sys­tem, mobile app, TV app and more so I am nev­er bored. There is always some­thing to fig­ure out in terms of what to build and how it should work and look. For the past cou­ple of years I was always too busy man­ag­ing the stu­dio to real­ly get into the details of design but now I can total­ly focus on that and it real­ly is a pleasure.

On the peo­ple side we have a small but grow­ing team of bril­liant indi­vid­u­als hal­ing from var­i­ous parts of the region includ­ing Viet­nam, Myan­mar and India. This lends an addi­tion­al lay­er of fun chal­lenge to the goings on as we con­stant­ly nego­ti­ate our dif­fer­ences but also dis­cov­er the many com­mon­al­i­ties afford­ed by the glob­alised tech indus­try. I also get to trav­el to Ho Chi Minh City reg­u­lar­ly which is a nice change from the extreme order that is Singapore. 

It is ear­ly days so I not only get to help shape the prod­uct from the very start but also the com­pa­ny itself. This includes fig­ur­ing out and main­tain­ing design and devel­op­ment process­es. For this I find my Boy­di­an explo­rations quite use­ful, paired with what is now more than 13 years of indus­try expe­ri­ence (how did that hap­pen?) I have also con­duct­ed more hir­ing inter­views in the past few months than I did in the ten years before.

In a month or two a first ver­sion of the prod­uct should be in the mar­ket. When we’ve got­ten to that point I will do anoth­er of these updates. In the mean­time just know I am up to my armpits in think­ing-through-mak­ing about art dis­cov­ery and enjoy­ment on screens small and large. If you have any­thing relat­ed to share, or would like to be one of the first to test-dri­ve the thing when it arrives, let me know.

Artificial intelligence, creativity and metis

Boris point­ed me to Cre­ativeAI, an inter­est­ing arti­cle about cre­ativ­i­ty and arti­fi­cial intel­li­gence. It offers a real­ly nice overview of the devel­op­ment of the idea of aug­ment­ing human capa­bil­i­ties through tech­nol­o­gy. One of the claims the authors make is that arti­fi­cial intel­li­gence is mak­ing cre­ativ­i­ty more acces­si­ble. Because tools with AI in them sup­port humans in a range of cre­ative tasks in a way that short­cuts the tra­di­tion­al require­ments of long prac­tice to acquire the nec­es­sary tech­ni­cal skills. 

For exam­ple, Shad­ow­Draw (PDF) is a pro­gram that helps peo­ple with free­hand draw­ing by guess­ing what they are try­ing to cre­ate and show­ing a dynam­i­cal­ly updat­ed ‘shad­ow image’ on the can­vas which peo­ple can use as a guide.

It is an inter­est­ing idea and in some ways these kinds of soft­ware indeed low­er the thresh­old for peo­ple to engage in cre­ative tasks. They are good exam­ples of arti­fi­cial intel­li­gence as part­ner in stead of mas­ter or servant. 

While read­ing Cre­ativeAI I wasn’t entire­ly com­fort­able though and I think it may have been caused by two things. 

One is that I care about cre­ativ­i­ty and I think that a good under­stand­ing of it and a dai­ly prac­tice at it—in the broad sense of the word—improves lives. I am also in some ways old-fash­ioned about it and I think the joy of cre­ativ­i­ty stems from the infi­nite­ly high skill ceil­ing involved and the nev­er-end­ing prac­tice it affords. Let’s call it the Jiro per­spec­tive, after the sushi chef made famous by a won­der­ful doc­u­men­tary.

So, claim­ing that cre­ative tools with AI in them can short­cut all of this life-long joy­ful toil pro­duces a degree of pan­ic for me. Although it’s prob­a­bly a Pas­toral world­view which would be bet­ter to aban­don. In a world eat­en by soft­ware, it’s bet­ter to be a Promethean.

The sec­ond rea­son might hold more water but real­ly is more of an open ques­tion than some­thing I have researched in any mean­ing­ful way. I think there is more to cre­ativ­i­ty than just the tech­ni­cal skill required and as such the Cre­ativeAI sto­ry runs the risk of being reduc­tion­ist. While read­ing the arti­cle I was also slow­ly but sure­ly mak­ing my way through one of the final chap­ters of James C. Scott’s See­ing Like a State, which is about the con­cept of metis.

It is prob­a­bly the most inter­est­ing chap­ter of the whole book. Scott intro­duces metis as a form of knowl­edge dif­fer­ent from that pro­duced by sci­ence. Here are some quick excerpts from the book that pro­vide a sense of what it is about. But I real­ly can’t do the rich­ness of his descrip­tion jus­tice here. I am try­ing to keep this short.

The kind of knowl­edge required in such endeav­ors is not deduc­tive knowl­edge from first prin­ci­ples but rather what Greeks of the clas­si­cal peri­od called metis, a con­cept to which we shall return. […] metis is bet­ter under­stood as the kind of knowl­edge that can be acquired only by long prac­tice at sim­i­lar but rarely iden­ti­cal tasks, which requires con­stant adap­ta­tion to chang­ing cir­cum­stances. […] It is to this kind of knowl­edge that [social­ist writer] Lux­em­burg appealed when she char­ac­ter­ized the build­ing of social­ism as “new ter­ri­to­ry” demand­ing “impro­vi­sa­tion” and “cre­ativ­i­ty.”

Scott’s argu­ment is about how author­i­tar­i­an high-mod­ernist schemes priv­i­lege sci­en­tif­ic knowl­edge over metis. His explo­ration of what metis means is super inter­est­ing to any­one ded­i­cat­ed to hon­ing a craft, or to cul­ti­vat­ing organ­i­sa­tions con­ducive to the devel­op­ment and appli­ca­tion of craft in the face of uncer­tain­ty. There is a close link between metis and the con­cept of agility.

So cir­cling back to arti­fi­cial­ly intel­li­gent tools for cre­ativ­i­ty I would be inter­est­ed in explor­ing not only how we can dimin­ish the need for the acqui­si­tion of the tech­ni­cal skills required, but to also accel­er­ate the acqui­si­tion of the prac­ti­cal knowl­edge required to apply such skills in the ever-chang­ing real world. I sug­gest we expand our under­stand­ing of what it means to be cre­ative, but with­out los­ing the link to actu­al practice.

For the ancient Greeks metis became syn­ony­mous with a kind of wis­dom and cun­ning best exem­pli­fied by such fig­ures as Odysseus and notably also Prometheus. The lat­ter in par­tic­u­lar exem­pli­fies the use of cre­ativ­i­ty towards trans­for­ma­tive ends. This is the real promise of AI for cre­ativ­i­ty in my eyes. Not to sim­ply make it eas­i­er to repro­duce things that used to be hard to cre­ate but to cre­ate new kinds of tools which have the capac­i­ty to sur­prise their users and to pro­duce results that were impos­si­ble to cre­ate before.